Normally Closed vs Normally Open Devices on One Zone

Why this matters

Almost every intrusion device is normally closed, and the reason is simple: a closed circuit that opens to alarm also alarms when someone cuts its wire. The handful of devices that are normally open (some panic buttons, pressure mats, a few dry-contact sensors) close to signal, which is backwards from everything around them. Put one of each on the same zone and you can build a loop that meters perfectly, passes a walk test, and still has a panic button that reports a trouble instead of an alarm, or goes dead the moment somebody bypasses an open window.

The two contact types, defined by their resting state

"Normally" means the device's state when nothing is happening: door shut, nobody in the room, button not pressed.

• Normally closed (NC): continuity at rest, opens to signal. Door and window contacts, most motion detector relays, glass-break relays, most tamper switches.
• Normally open (NO): no continuity at rest, closes to signal. Many panic and hold-up buttons, pressure mats, some water sensors' dry contacts, and the detectors on conventional two-wire fire initiating circuits, which signal by drawing the loop down rather than opening it.

Why intrusion settled on NC. Strip the resistor off both loops and cut the wire. The NC loop reads open, exactly as if a door opened, so an armed panel alarms: the failure announces itself. The NO loop reads open too, but open is its resting state, so the panel sees nothing at all and the device is dead from that moment on. An intruder who finds the cable has a reason to cut it on an NO loop and a reason to leave it alone on an NC one. The end-of-line resistor narrows the gap (it lets an NO loop report a cut as an open), but the habit, the parts on the truck and the panels' default zone responses were all built around NC first.

Many devices with a relay offer both. A form C relay has three terminals: common, NC and NO. You choose the behavior by which two terminals you land. On an intrusion zone, land common and NC unless the documentation for the panel and zone type says otherwise. Plenty of panic buttons ship with a jumper or switch to select NC or NO; set it, then check it with a meter rather than trusting the label.

How each one is wired against the end-of-line resistor

The resistor sits at the last device either way (see related: End-of-Line Resistor Supervision: Open, Short and Normal). What changes is how the devices sit relative to the pair:

PANEL       NC DEVICES IN SERIES ON ONE CONDUCTOR
Z4  o---[W1 NC]---[W2 NC]---+---------+---------+
                            |         |         |
                        [PB NO]   [MAT NO]  [EOL R]
                            |         |      at end
COM o-----------------------+---------+---------+
            NO DEVICES IN PARALLEL ACROSS THE PAIR

W1 or W2 opens    -> loop reads open
PB or MAT closes  -> loop reads short (wire ohms)
all at rest       -> loop reads EOL value

An NC device in series opens the path: the panel reads open. An NO device across the pair closes and shunts the resistor: the panel reads a short. So a zone built like this signals by TWO different faults, and it only works as an alarm zone if the panel treats both open and short as alarm for that zone type. That is the first thing to check, and it is the first place this design breaks.

The two wiring mistakes you will see right away, and the two you will not

Visible on the first meter reading:

• An NO device wired in series. At rest it is open, so the loop reads open forever. The zone will not arm. Nobody ships that.
• An NC device wired across the pair. At rest it is closed, so it shorts the loop forever. Constant short trouble. Also caught on day one.

Invisible until the day it matters:

• The panel reports a short as trouble, not alarm. Plenty of panels, on a single-EOL burglary zone, treat a short as a trouble or tamper condition, per their documentation. An NO panic button across that loop "works" in the sense that the keypad shows something when it is pressed. It does not send an alarm.
• An open NC device upstream disconnects the NO devices downstream. Look at the diagram. If W2 is open, the path from the panel to PB, MAT and the resistor is broken at W2. Press PB: nothing reaches the panel, because the panic button is now on a dead stub. A window left open on a summer night disables the panic button beside the bed.

And one that is not wiring at all: the zone type is per zone, not per device. A panic button riding on a perimeter zone inherits that zone's behavior. Bypass the zone because a window is open and the panic is bypassed with it. Put the same button on an entry/exit zone and pressing it starts an entry delay. A burglary report from zone 4 also tells the central station "intrusion, bedroom windows", not "hold-up", and the operator follows a burglary script for what the customer meant as a life-safety press.

A takeover zone, traced

A house comes over from another dealer. Zone 4 is labeled "MBR" and contains two bedroom window contacts (W1, W2, NC in series) and a bedside panic button (PB, NO, across the pair), with the resistor at the panic button box. The previous installer saved a zone input. The panel's documentation says a short on this zone type is reported as a trouble while disarmed and as an alarm while armed. Zone 4 is perimeter instant.

Account on test first; see related: How to Put an Alarm Account on Test and Take It Off. Lift zone 4 at the panel to meter it.

Metered at the panel, zone lifted. All at rest: the EOL value. Open W1: open. Close W1, press PB: about 1.3 ohms, which is 80 ft of round-trip 22 AWG conductor (40 ft of cable to the bedroom, 2 x 40 x 16.1 / 1,000). Every reading is "correct". The previous tech almost certainly stopped here.

Now run the three cases the customer will actually live through:

• Disarmed, afternoon, customer hears someone at the back door and presses PB. Short, disarmed: trouble. The keypad beeps. No alarm signal leaves the house.
• Armed Stay, W2 left open for air, zone 4 bypassed at arming. PB is bypassed with it. Even if it were not, W2 open has disconnected PB from the panel. Press it: nothing.
• Armed Away, all closed, PB pressed. Short, armed: alarm. The central station gets a burglary on zone 4, and dispatches on a burglary.

One of three works, and the one that works does so under the wrong event type. The customer bought a panic button for the first two situations.

The fix. Pull PB off zone 4. Move the EOL resistor to W2, which is now the last device on the window loop. Put PB on its own zone programmed 24-hour panic (audible or silent, the customer's choice, written on the zone sheet), either a hardwired expander zone on a separate pair or a wireless panic enrolled to the panel. Re-meter zone 4: EOL value at rest, open with either window open (the resistor sits beyond both contacts), and a jumper across the pair inside W2's box drops it to wire ohms (the pair really runs to that box).

Retest the three cases. Disarmed, PB pressed: panic alarm. Stay with W2 bypassed: PB still alarms, because its zone is not affected by zone 4's bypass. Away: panic alarm, reported as a panic, not a burglary. Confirm every event at the central station with the correct zone and event code before taking the account off test.

When a mixed loop is defensible

A mixed NC and NO loop is not forbidden. Some panels document it and treat open and short both as alarm on certain zone types. It can be reasonable where:

• every device on the loop is the same function (all intrusion, say an NC door contact and an NO pressure mat inside the same door), so sharing a zone type is honest;
• the panel's documentation states short is alarm for that zone type in every arm state you will use;
• the NO devices sit downstream of nothing that a customer will routinely leave open.

It is never defensible for a life-safety or hold-up device sharing a zone with intrusion devices. Those get their own zone because their zone TYPE is different, not because the wiring will not carry them.

Where normally open still has the better case

Fire initiating circuits are the big one: conventional detectors are designed to signal by changing the loop's current, and the panel is designed to read it that way, under NFPA 72 in the edition your jurisdiction has adopted. That is fire work, often separately licensed; see related: Combination Fire and Burglary Panels and the License Line.

On the intrusion side, an NO device is the honest choice when the device itself only offers it and the zone is dedicated to it: an older pressure mat on its own zone, or a dry-contact water sensor on its own 24-hour non-burglary zone. With an EOL across the end, even an NO loop is supervised for an open wire, which an unterminated NO loop is not: cut it and nothing ever reads different.

How to verify you got this right

On every zone, meter three readings at the panel with the zone lifted and write them on the zone sheet: at rest (EOL value plus wire), each NC device opened (open), each NO device operated (wire ohms). Then for every zone that carries an NO device, test it at the keypad in Off, Stay and Away, and with any other device on that zone opened or bypassed. The second half of that sentence is the test that catches the takeover above, and it is the one nobody runs.

References

• Panel manufacturer's installation manual (open and short response by zone type and arm state)
• Device manufacturers' instructions (form C terminals, NC/NO selection)
• NFPA 72, National Fire Alarm and Signaling Code, in the edition adopted by your jurisdiction
• See related: End-of-Line Resistor Supervision: Open, Short and Normal; Zone Types and What Each One Does When Tripped; Security + Alarm Systems Reference